Objective responses in patients with malignant melanoma or renal cell cancer in early clinical studies do not predict regulatory approval

Purpose: Tumor responses in early-phase trials are used to determine whether new agents warrant further study. Given that spontaneous regressions are observed in melanoma and renal cell carcinoma, this study assessed whether tumor responses, particularly in these two tumor types, predict for future regulatory drug approval. Experimental Design: The literature was reviewed to assess tumor response rates to cytotoxic agents in phase I and IIrials in the following solid tumors: melanoma, renal cell carcinoma, non small-cell lung cancer, breast cancer, ovarian cancer, colorectal cancer, and other solid tumors. Response rates were categorized and the relationship of these categories to the end point of regulatory drug approval was determined. Results: Fifty-eight drugs were assessed in 100 phase I trials, and 46 of these drugs were also studied in 499 phase IIrials. Higher overall response rates in both phase I trials (P = 0.03) and phase II trials (P < 0.0001) were predictive of regulatory approval. However, response in melanoma or renal cell carcinoma was not predictive for either phase I or phase IItudies. Conclusions: For cytotoxic agents, although overall objective response rates reliably predict subsequent marketing approval, isolated responses in melanoma and renal cell carcinoma are not predictive.